/*********************************************************************
*
* Func Name : GetCharCount
*
* Description: Get the count of char which is in sWord and in sCharSet
*
* Parameters : sWord: the word
*
* Returns : COUNT
* Author : Kevin Zhang
* History :
* 1.create 2002-5-21
2. corrected by Wang Zhifu 2004-2-18
*********************************************************************/
int GetCharCount(char *sCharSet,char *sWord)
{
unsigned int k=0;
char tchar[3];
int nCount=0;
tchar[2]=0;
while(k < strlen(sWord))
{
tchar[0]=sWord[k];
tchar[1]=0;
if(sWord[k]<0)
{
tchar[1]=sWord[k+1];
k+=1;
}
k+=1;
// 这里有改动，原语句为
// if((tchar[0]<0&&CC_Find(sCharSet, tchar))||strchr(sCharSet,tchar[0]))
// 改为：因为后面部分经常为true!!!应该加入限制条件！
// corrected by Wang Zhifu
if(((tchar[0]<0&&(CC_Find(sCharSet, tchar)!=NULL)))||(tchar[0]>0&&(strchr(sCharSet,tchar[0])!=NULL)))
nCount++;
}
return nCount;
}
bool PostfixSplit(char *sWord, char *sWordRet, char *sPostfix)
{
char sSinglePostfix[]=POSTFIX_SINGLE;
char sMultiPostfix[][9]=POSTFIX_MUTIPLE;
unsigned int nPostfixLen=0,nWordLen=strlen(sWord);
int i=0;
while(sMultiPostfix[i][0]!=0&&strncmp(sWord+nWordLen-strlen(sMultiPostfix[i]),sMultiPostfix[i],strlen(sMultiPostfix[i]))!=0)
{//Try to get the postfix of an address
i++;
}
strcpy(sPostfix,sMultiPostfix[i]);
nPostfixLen=strlen(sMultiPostfix[i]);//Get the length of place postfix
if(nPostfixLen==0)
{
sPostfix[2]=0;
strncpy(sPostfix,sWord+nWordLen-2,2);
if(CC_Find(sSinglePostfix,sPostfix))
nPostfixLen=2;
}
strncpy(sWordRet,sWord,nWordLen-nPostfixLen);
sWordRet[nWordLen-nPostfixLen]=0;//Get the place name which have erasing the postfix
sPostfix[nPostfixLen]=0;
return true;
}
/*********************************************************************
*
* Func Name : GetMaxMatch
*
* Description: Get the max match to the word
*
*
* Parameters : nHandle: the only handle which will be attached to the word
* Returns : success or fail
* Author : Kevin Zhang
* History :
* 1.create 2002-1-21
*********************************************************************/
bool CDictionary::GetMaxMatch(char *sWord, char *sWordRet,int *npHandleRet)
{
char sWordGet[WORD_MAXLENGTH-2],sFirstChar[3];
int nPos,nFoundPos,nTemp;
PWORD_CHAIN pCur;
*npHandleRet=-1;
if(!PreProcessing(sWord, &nPos,sWordGet))
return false;
if (nPos < 0) return false;
sWordRet[0]=0;
strncpy(sFirstChar,sWord,strlen(sWord)-strlen(sWordGet));//Get the first char
sFirstChar[strlen(sWord)-strlen(sWordGet)]=0;//Set the end flag
FindInOriginalTable(nPos,sWordGet,-1,&nFoundPos);
nTemp=nFoundPos;//Check its previous position
if(nFoundPos==-1)
nTemp=0;
while(nTemp<m_IndexTable[nPos].nCount&&CC_Find(m_IndexTable[nPos].pWordItemHead[nTemp].sWord,sWordGet)!=m_IndexTable[nPos].pWordItemHead[nTemp].sWord)
{//Get the next
nTemp+=1;
}
if(nTemp<m_IndexTable[nPos].nCount&&CC_Find(m_IndexTable[nPos].pWordItemHead[nTemp].sWord,sWordGet)==m_IndexTable[nPos].pWordItemHead[nTemp].sWord)
{
strcpy(sWordRet,sFirstChar);
strcat(sWordRet,m_IndexTable[nPos].pWordItemHead[nTemp].sWord);
*npHandleRet=m_IndexTable[nPos].pWordItemHead[nTemp].nHandle;
return true;
}//Cannot get the item and retrieve the modified data if exists
//Operation in the index table and its items
if(m_pModifyTable&&m_pModifyTable[nPos].pWordItemHead)//Exists
pCur=m_pModifyTable[nPos].pWordItemHead;
else
pCur=NULL;
while(pCur!=NULL&&strcmp(pCur->data.sWord,sWordGet)<=0&&CC_Find(pCur->data.sWord,sWordGet)!=pCur->data.sWord)//
{
pCur=pCur->next;
}
if(pCur!=NULL&&CC_Find(pCur->data.sWord,sWordGet)!=pCur->data.sWord)
{//Get it
strcpy(sWordRet,sFirstChar);
strcat(sWordRet,pCur->data.sWord);
*npHandleRet=pCur->data.nHandle;
return true;
}
return false;
}
/*********************************************************************
*
* Func Name : IsForeign
*
* Description: Decide whether the word is Chinese Num word
*
* Parameters : sWord: the word
*
* Returns : the index value
* Author : Kevin Zhang
* History :
* 1.create 2002-1-26
*********************************************************************/
bool IsAllChineseNum(char *sWord)
{//百分之五点六的人早上八点十八分起床
unsigned int k;
char tchar[3];
char ChineseNum[]="零○一二两三四五六七八九十廿百千万亿壹贰叁肆伍陆柒捌玖拾佰仟∶·．／点";//
char sPrefix[]="几数第上成";
for(k = 0; k < strlen(sWord); k+=2)
{
strncpy(tchar,sWord+k,2) ;
tchar[2]='\0';
if(strncmp(sWord+k,"分之",4)==0)//百分之五
{
k+=2;
continue;
}
if(!CC_Find(ChineseNum, tchar)&&!(k==0&&CC_Find(sPrefix, tchar)))
return false;
}
return true;
}
bool CSegment::IsYearTime(char *sNum)
{
//Judge whether the sNum is a num genearating year
unsigned int nLen=strlen(sNum);
char sTemp[3];
strncpy(sTemp,sNum,2);
sTemp[2]=0;
if(IsAllSingleByte((unsigned char *)sNum)&&(nLen>=3||nLen==2&&sNum[0]>'4'))//1992年, 90年
return true;
if(IsAllNum((unsigned char *)sNum)&&(nLen>=6||nLen==4&&CC_Find("５６７８９",sTemp)))
return true;
if(GetCharCount("零○一二三四五六七八九壹贰叁肆伍陆柒捌玖",sNum)==(int)nLen/2&&nLen>=3)
return true;
if(nLen==8&&GetCharCount("千仟零○",sNum)==2)//二仟零二年
return true;
return false;
}
/*********************************************************************
*
* Func Name : GetCharCount
*
* Description: Get the count of char which is in sWord and in sCharSet
*
* Parameters : sWord: the word
*
* Returns : COUNT
* Author : Kevin Zhang
* History :
* 1.create 2002-5-21
*********************************************************************/
int GetCharCount(char *sCharSet,char *sWord)
{
unsigned int k=0;
char tchar[3];
int nCount=0;
tchar[2]=0;
while(k < strlen(sWord))
{
tchar[0]=sWord[k];
tchar[1]=0;
if(sWord[k]<0)
{
tchar[1]=sWord[k+1];
k+=1;
}
k+=1;
if((tchar[0]<0&&CC_Find(sCharSet, tchar))||strchr(sCharSet,tchar[0]))
nCount++;
}
return nCount;
}
/*********************************************************************
*
* Func Name : IsNumExist
*
* Description: Judge whether there is Num Char in the string
*
*
* Parameters : sSentence: the original sentence which includes Chinese or Non-Chinese char
*
* Returns : the end of the sub-sentence
* Author : Wangzhifu
* History :
* 1.create 2004-3-2
*********************************************************************/
bool IsNumExist(char *sWord)
{
unsigned int k;
char tchar[3];
char ChineseNum[]="零○一二两三四五六七八九十廿百千万亿壹贰叁肆伍陆柒捌玖拾佰仟";//
for(k = 0; k < strlen(sWord); k++)
{
tchar[0]=sWord[k];
if(tchar[0]<0)
{
tchar[1]=sWord[k++];
tchar[2]='\0';
if(CC_Find(ChineseNum, tchar))
return true;
}
else
{
if((tchar[0]>'0'-1)&&(tchar[0]<'9'+1))
return true;
}
}
return false;
}
bool CSegGraph::GenerateWordNet(char *sSentence,CDictionary &dictCore,bool bOriginalFreq)
{
//Gernerate the word net from the sLine, that's list all the possible word
unsigned int i=0,j,nLen=strlen(sSentence);
char sWord[WORD_MAXLENGTH]="",sTempWord[WORD_MAXLENGTH]="",sWordMatch[WORD_MAXLENGTH];
int nWordIndex=0,nHandleTemp,k,nPOS;
int nMatchFreq[20],nMatchHandle[20],nTotalFreq,nMatchCount;
double dValue=0;
m_nAtomCount=0;
m_segGraph.SetEmpty();//Set segmentation graph empty
AtomSegment(sSentence);
//Atomic Segmentation
for(i=0;i<m_nAtomCount;i++)//Init the cost array
{
if(m_nAtomPOS[i]==CT_CHINESE)//The atom is a Chinese Char
{
if(!bOriginalFreq)//Not original frequency
m_segGraph.SetElement(i,i+1,log(MAX_FREQUENCE),0);//init the link with the maximum value
else
m_segGraph.SetElement(i,i+1,0,0,m_sAtom[i]);//init the link with the maximum value
}
else//Other atom
{
strcpy(sWord,m_sAtom[i]);//init the word
dValue=MAX_FREQUENCE;
switch(m_nAtomPOS[i])
{
case CT_INDEX:
case CT_NUM:
nPOS=-27904;//'m'*256
strcpy(sWord,"未##数");
dValue=0;
break;
case CT_DELIMITER:
nPOS=30464;//'w'*256;
break;
case CT_LETTER:
nPOS=-'n'*256-'x';//
dValue=0;
strcpy(sWord,"未##串");
break;
case CT_SINGLE://12021-2129-3121
if(GetCharCount("+-1234567890",m_sAtom[i])==(int)strlen(m_sAtom[i]))
{
nPOS=-27904;//'m'*256
strcpy(sWord,"未##数");
}
else
{
nPOS=-'n'*256-'x';//
strcpy(sWord,"未##串");
}
dValue=0;
break;
default:
nPOS=m_nAtomPOS[i];//'?'*256;
break;
}
if(!bOriginalFreq)//Not original frequency
m_segGraph.SetElement(i,i+1,0,nPOS);//init the link with minimum
else
m_segGraph.SetElement(i,i+1,dValue,nPOS,sWord);//init the link with minimum
}
}
i=0;
while(i<m_nAtomCount)//All the word
{
strcpy(sWord,m_sAtom[i]);//Get the current atom
j=i+1;
if(strcmp(sWord,"月")==0&&strcmp(m_sAtom[i+1],"份")==0)//Don't split 月份
j+=1;
while(j<=m_nAtomCount&&dictCore.GetMaxMatch(sWord,sWordMatch,&nHandleTemp))
{//Add a condition to control the end of string
//retrieve the dictionary with the word
if(strcmp(sWordMatch,sWord)==0)//find the current word
{
nTotalFreq=0;
dictCore.GetHandle(sWord,&nMatchCount,nMatchHandle,nMatchFreq);
for(k=0;k<nMatchCount;k++)//Add the frequency
{
nTotalFreq+=nMatchFreq[k];
}
//Adding a rule to exclude some words to be formed.
if(strlen(sWord)==4&&i>=1&&(IsAllNum((unsigned char *)m_sAtom[i-1])||IsAllChineseNum(m_sAtom[i-1]))&&(strncmp(sWord,"年",2)==0||strncmp(sWord,"月",2)==0))
{//1年内、1999年末
if(CC_Find("末内中底前间初",sWord+2))
break;
}
if(nMatchCount==1)//The possible word has only one POS, store it
{
if(!bOriginalFreq)//Not original frequency
m_segGraph.SetElement(i,j,-log(nTotalFreq+1)+log(MAX_FREQUENCE),nMatchHandle[0]);
else
m_segGraph.SetElement(i,j,nTotalFreq,nMatchHandle[0],sWord);
}
else
{
if(!bOriginalFreq)//Not original frequency
m_segGraph.SetElement(i,j,-log(nTotalFreq+1)+log(MAX_FREQUENCE),0);
else
m_segGraph.SetElement(i,j,nTotalFreq,0,sWord);
}
}
strcat(sWord,m_sAtom[j++]);
}
i+=1;//Start from i++;
}
return true;
}
//Guess the POS of No. nIndex word item
bool CSpan::GuessPOS(int nIndex,int *pSubIndex)
{
int j=0,i=nIndex,nCharType;
unsigned int nLen;
switch(m_tagType)
{
case TT_NORMAL:
break;
case TT_PERSON:
j=0;
if(CC_Find("××",m_sWords[nIndex]))
{
m_nTags[i][j]=6;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,6)+1);
}
else
{
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
nLen=strlen(m_sWords[nIndex]);
if(nLen>=4)
{
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)*8);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*8);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*8);
}
else if(nLen==2)
{
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
nCharType=charType((unsigned char *)m_sWords[nIndex]);
if(nCharType==CT_OTHER||nCharType==CT_CHINESE)
{
m_nTags[i][j]=1;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,1)+1);
m_nTags[i][j]=2;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,2)+1);
m_nTags[i][j]=3;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,3)+1);
m_nTags[i][j]=4;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,4)+1);
}
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)*8);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*8);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*8);
}
}
break;
case TT_PLACE:
j=0;
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
nLen=strlen(m_sWords[nIndex]);
if(nLen>=4)
{
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)*8);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*8);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*8);
}
else if(nLen==2)
{
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
nCharType=charType((unsigned char *)m_sWords[nIndex]);
if(nCharType==CT_OTHER||nCharType==CT_CHINESE)
{
m_nTags[i][j]=1;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,1)+1);
m_nTags[i][j]=2;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,2)+1);
m_nTags[i][j]=3;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,3)+1);
m_nTags[i][j]=4;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,4)+1);
}
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)*8);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*8);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*8);
}
break;
case TT_TRANS:
j=0;
nLen=strlen(m_sWords[nIndex]);
m_nTags[i][j]=0;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,0)+1);
if(!IsAllChinese((unsigned char *)m_sWords[nIndex]))
{
if(IsAllLetter((unsigned char *)m_sWords[nIndex]))
{
m_nTags[i][j]=1;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,1)+1);
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)+1);
/* }
if(IsAllNum((unsigned char *)m_sWords[nIndex])||IsAllLetter((unsigned char *)m_sWords[nIndex]))
{
*/ m_nTags[i][j]=2;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,2)*2+1);
m_nTags[i][j]=3;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,3)*2+1);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*2+1);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*2+1);
}
m_nTags[i][j]=41;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,41)*8);
m_nTags[i][j]=42;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,42)*8);
m_nTags[i][j]=43;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,43)*8);
}
else if(nLen>=4)
{
m_nTags[i][j]=41;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,41)*8);
m_nTags[i][j]=42;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,42)*8);
m_nTags[i][j]=43;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,43)*8);
}
else if(nLen==2)
{
nCharType=charType((unsigned char *)m_sWords[nIndex]);
if(nCharType==CT_OTHER||nCharType==CT_CHINESE)
{
m_nTags[i][j]=1;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,1)*2+1);
m_nTags[i][j]=2;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,2)*2+1);
m_nTags[i][j]=3;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,3)*2+1);
m_nTags[i][j]=30;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,30)*8+1);
m_nTags[i][j]=11;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,11)*4+1);
m_nTags[i][j]=12;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,12)*4+1);
m_nTags[i][j]=13;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,13)*4+1);
m_nTags[i][j]=21;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,21)*2+1);
m_nTags[i][j]=22;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,22)*2+1);
m_nTags[i][j]=23;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,23)*2+1);
}
m_nTags[i][j]=41;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,41)*8);
m_nTags[i][j]=42;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,42)*8);
m_nTags[i][j]=43;
m_dFrequency[i][j++]=(double)1/(double)(m_context.GetFrequency(0,43)*8);
}
break;
default:
break;
}
*pSubIndex=j;
return true;
}
//Generate Word according the segmentation route
bool CSegment::GenerateWord(int **nSegRoute, int nIndex)
{
unsigned int i=0,k=0;
int j,nStartVertex,nEndVertex,nPOS;
char sAtom[WORD_MAXLENGTH],sNumCandidate[100],sCurWord[100];
ELEMENT_TYPE fValue;
while(nSegRoute[nIndex][i]!=-1&&nSegRoute[nIndex][i+1]!=-1&&nSegRoute[nIndex][i]<nSegRoute[nIndex][i+1])
{
nStartVertex=nSegRoute[nIndex][i];
j=nStartVertex;//Set the start vertex
nEndVertex=nSegRoute[nIndex][i+1];//Set the end vertex
nPOS=0;
m_graphSeg.m_segGraph.GetElement(nStartVertex,nEndVertex,&fValue,&nPOS);
sAtom[0]=0;
while(j<nEndVertex)
{//Generate the word according the segmentation route
strcat(sAtom,m_graphSeg.m_sAtom[j]);
j++;
}
m_pWordSeg[nIndex][k].sWord[0]=0;//Init the result ending
strcpy(sNumCandidate,sAtom);
while(sAtom[0]!=0&&(IsAllNum((unsigned char *)sNumCandidate)||IsAllChineseNum(sNumCandidate)))
{//Merge all seperate continue num into one number
//sAtom[0]!=0: add in 2002-5-9
strcpy(m_pWordSeg[nIndex][k].sWord,sNumCandidate);
//Save them in the result segmentation
i++;//Skip to next atom now
sAtom[0]=0;
while(j<nSegRoute[nIndex][i+1])
{//Generate the word according the segmentation route
strcat(sAtom,m_graphSeg.m_sAtom[j]);
j++;
}
//add a judge for out of memery,
//del, not nessasseray for ordinary text file,becasuse no word's lenth can larger than100
//so remain the same
// if(strlen(sNumCandidate)+strlen(sAtom)<100)
strcat(sNumCandidate,sAtom);
// else
// break;
}
unsigned int nLen=strlen(m_pWordSeg[nIndex][k].sWord);
if(nLen==4&&CC_Find("第上成±—＋∶·．／",m_pWordSeg[nIndex][k].sWord)||nLen==1&&strchr("+-./",m_pWordSeg[nIndex][k].sWord[0]))
{//Only one word
strcpy(sCurWord,m_pWordSeg[nIndex][k].sWord);//Record current word
i--;
}
else if(m_pWordSeg[nIndex][k].sWord[0]==0)//Have never entering the while loop
{
strcpy(m_pWordSeg[nIndex][k].sWord,sAtom);
//Save them in the result segmentation
strcpy(sCurWord,sAtom);//Record current word
}
else
{//It is a num
if(strcmp("－－",m_pWordSeg[nIndex][k].sWord)==0||strcmp("—",m_pWordSeg[nIndex][k].sWord)==0||m_pWordSeg[nIndex][k].sWord[0]=='-'&&m_pWordSeg[nIndex][k].sWord[1]==0)//The delimiter "－－"
{
nPOS=30464;//'w'*256;Set the POS with 'w'
i--;//Not num, back to previous word
}
else
{//Adding time suffix
char sInitChar[3];
unsigned int nCharIndex=0;//Get first char
sInitChar[nCharIndex]=m_pWordSeg[nIndex][k].sWord[nCharIndex];
if(sInitChar[nCharIndex]<0)
{
nCharIndex+=1;
sInitChar[nCharIndex]=m_pWordSeg[nIndex][k].sWord[nCharIndex];
}
nCharIndex+=1;
sInitChar[nCharIndex]='\0';
if(k>0&&(abs(m_pWordSeg[nIndex][k-1].nHandle)==27904||abs(m_pWordSeg[nIndex][k-1].nHandle)==29696)&&(strcmp(sInitChar,"—")==0||sInitChar[0]=='-')&&(strlen(m_pWordSeg[nIndex][k].sWord)>nCharIndex))
{//3-4月 //27904='m'*256
//Split the sInitChar from the original word
strcpy(m_pWordSeg[nIndex][k+1].sWord,m_pWordSeg[nIndex][k].sWord+nCharIndex);
m_pWordSeg[nIndex][k+1].dValue=m_pWordSeg[nIndex][k].dValue;
m_pWordSeg[nIndex][k+1].nHandle=27904;
m_pWordSeg[nIndex][k].sWord[nCharIndex]=0;
m_pWordSeg[nIndex][k].dValue=0;
m_pWordSeg[nIndex][k].nHandle=30464;//'w'*256;
m_graphOptimum.SetElement(nStartVertex,nStartVertex+1,m_pWordSeg[nIndex][k].dValue,m_pWordSeg[nIndex][k].nHandle,m_pWordSeg[nIndex][k].sWord);
nStartVertex+=1;
k+=1;
}
nLen=strlen(m_pWordSeg[nIndex][k].sWord);
if((strlen(sAtom)==2&&CC_Find("月日时分秒",sAtom))||strcmp(sAtom,"月份")==0)
{//2001年
strcat(m_pWordSeg[nIndex][k].sWord,sAtom);
strcpy(sCurWord,"未##时");
nPOS=-29696;//'t'*256;//Set the POS with 'm'
}
else if(strcmp(sAtom,"年")==0)
{
if(IsYearTime(m_pWordSeg[nIndex][k].sWord))//strncmp(sAtom,"年",2)==0&&
{//1998年，
strcat(m_pWordSeg[nIndex][k].sWord,sAtom);
strcpy(sCurWord,"未##时");
nPOS=-29696;//Set the POS with 't'
}
else
{
strcpy(sCurWord,"未##数");
nPOS=-27904;//Set the POS with 'm'
i--;//Can not be a time word
}
}
else
{
//早晨/t 五点/t
if(strcmp(m_pWordSeg[nIndex][k].sWord+strlen(m_pWordSeg[nIndex][k].sWord)-2,"点")==0)
{
strcpy(sCurWord,"未##时");
nPOS=-29696;//Set the POS with 't'
}
else
{
// if(!CC_Find("∶·．／",m_pWordSeg[nIndex][k].sWord+nLen-2)&&m_pWordSeg[nIndex][k].sWord[nLen-1]!='.'&&m_pWordSeg[nIndex][k].sWord[nLen-1]!='/')
// {
/// strcpy(sCurWord,"未##数");
// nPOS=-27904;//'m'*256;Set the POS with 'm'
// }
// else if(nLen>strlen(sInitChar))
// {//Get rid of . example 1.
// if(m_pWordSeg[nIndex][k].sWord[nLen-1]=='.'||m_pWordSeg[nIndex][k].sWord[nLen-1]=='/')
// m_pWordSeg[nIndex][k].sWord[nLen-1]=0;
// else
// m_pWordSeg[nIndex][k].sWord[nLen-2]=0;
// strcpy(sCurWord,"未##数");
// nPOS=-27904;//'m'*256;Set the POS with 'm'
// i--;
// }
//here's bug in it ＋．．．．．． do not del ．
//2004-3-2 fixed by Wangzhifu
// if(!CC_Find("∶·．／",m_pWordSeg[nIndex][k].sWord+nLen-2)&&m_pWordSeg[nIndex][k].sWord[nLen-1]!='.'&&m_pWordSeg[nIndex][k].sWord[nLen-1]!='/')
//2004_06_18 修改,见04_06_18新华网语料bug.txt
if(!CC_Find("∶·．／",m_pWordSeg[nIndex][k].sWord+nLen-2)&&m_pWordSeg[nIndex][k].sWord[nLen-1]!='.')
{
strcpy(sCurWord,"未##数");
nPOS=-27904;//'m'*256;Set the POS with 'm'
}
else if(nLen>strlen(sInitChar))
{//Get rid of . example 1.
//,but if" ＋．．．．．．" do not del ．//fixed at 2004-3-2 by Wang Zhifu
char TempWord[100];
strcpy(TempWord,m_pWordSeg[nIndex][k].sWord);
TempWord[nLen-2]=0;//去掉最后一个字符！
if(IsNumExist(TempWord))
{
if(m_pWordSeg[nIndex][k].sWord[nLen-1]=='.'||m_pWordSeg[nIndex][k].sWord[nLen-1]=='/')
m_pWordSeg[nIndex][k].sWord[nLen-1]=0;
else
m_pWordSeg[nIndex][k].sWord[nLen-2]=0;
strcpy(sCurWord,"未##数");
nPOS=-27904;//'m'*256;Set the POS with 'm'
i--;
}
else
{
nPOS=-'n'*256-'x';//
strcpy(sCurWord,"未##串");
}
}
}
// end of fixed line;
i--;//Not num, back to previous word
}
}
fValue=0;
nEndVertex=nSegRoute[nIndex][i+1];//Ending POS changed to latter
}
m_pWordSeg[nIndex][k].nHandle=nPOS;//Get the POS of current word
m_pWordSeg[nIndex][k].dValue=fValue;//(int)(MAX_FREQUENCE*exp(-fValue));//Return the frequency of current word
m_graphOptimum.SetElement(nStartVertex,nEndVertex,fValue,nPOS,sCurWord);
//Generate optimum segmentation graph according the segmentation result
i++;//Skip to next atom
k++;//Accept next word
}
m_pWordSeg[nIndex][k].sWord[0]=0;
m_pWordSeg[nIndex][k].nHandle=-1;//Set ending
return true;
}
/*********************************************************************
*
* Func Name : IsAllNum
*
* Description: Judge the string is all made up of Num Char
*
*
* Parameters : sSentence: the original sentence which includes Chinese or Non-Chinese char
*
* Returns : the end of the sub-sentence
* Author : Kevin Zhang
* History :
* 1.create 2002-1-24
*********************************************************************/
bool IsAllNum(unsigned char *sString)
{
unsigned int nLen=strlen((const char *)sString),i=0;
char sChar[3];
sChar[2]=0;
if(i<nLen)//Get prefix such as + -
{
sChar[0]=sString[i++];
if(sChar[0]<0)//Get first char
sChar[1]=sString[i++];
else
sChar[1]=0;
if(!strstr("±+—-＋",sChar))
{
i=0;
}
}
while(i<nLen-1&&sString[i]==163&&sString[i+1]>175&&sString[i+1]<186)
{
i+=2;
}
if(i<nLen)//Get middle delimiter such as .
{
sChar[0]=sString[i++];
if(sChar[0]<0)//Get first char
sChar[1]=sString[i++];
else
sChar[1]=0;
if(CC_Find("∶·．／",sChar)||sChar[0]=='.'||sChar[0]=='/')
{//98．1％
while(i<nLen-1&&sString[i]==163&&sString[i+1]>175&&sString[i+1]<186)
{
i+=2;
}
}
else
{
i-=strlen(sChar);
}
}
if(i>=nLen)
return true;
while(i<nLen&&sString[i]>'0'-1&&sString[i]<'9'+1)
{//single byte number char
i+=1;
}
if(i<nLen)//Get middle delimiter such as .
{
sChar[0]=sString[i++];
if(sChar[0]<0)//Get first char
sChar[1]=sString[i++];
else
sChar[1]=0;
if(CC_Find("∶·．／",sChar)||sChar[0]=='.'||sChar[0]=='/')
{//98．1％
while(i<nLen&&sString[i]>'0'-1&&sString[i]<'9'+1)
{
i+=1;
}
}
else
{
i-=strlen(sChar);
}
}
if(i<nLen)//Get middle delimiter such as .
{
sChar[0]=sString[i++];
if(sChar[0]<0)//Get first char
sChar[1]=sString[i++];
else
sChar[1]=0;
if(!CC_Find("百千万亿佰仟％‰",sChar)&&sChar[0]!='%')
i-=strlen(sChar);
}
if(i>=nLen)
return true;
return false;
}
//Paragraph Segment and POS Tagging
bool CResult::ParagraphProcessing(char *sParagraph,char *sResult)
{
char *sSentence,sChar[3];
char *sSentenceResult;
unsigned int nLen=strlen(sParagraph)+13;
sSentence=new char[nLen];//malloc buffer
sSentenceResult=new char[nLen*3];//malloc buffer
sSentence[0]=0;
unsigned int nPosIndex=0,nParagraphLen=strlen(sParagraph),nSentenceIndex=0;
sChar[2]=0;
sResult[0]=0;//Init the result
bool bFirstIgnore=true;
strcpy(sSentence,SENTENCE_BEGIN);//Add a sentence begin flag
while(nPosIndex<nParagraphLen)
{//Find a whole sentence which separated by ! . \n \r
sChar[0]=sParagraph[nPosIndex];//Get a char
sChar[1]=0;
if(sParagraph[nPosIndex]<0)
{//double byte char
nPosIndex+=1;
sChar[1]=sParagraph[nPosIndex];
}
nPosIndex+=1;
/*
#define SEPERATOR_C_SENTENCE "。！？：；…"
#define SEPERATOR_C_SUB_SENTENCE "、，（）“”‘’"
#define SEPERATOR_E_SENTENCE "!?:;"
#define SEPERATOR_E_SUB_SENTENCE ",()\042'"
#define SEPERATOR_LINK "\n\r "
*/
if(CC_Find(SEPERATOR_C_SENTENCE,sChar)||CC_Find(SEPERATOR_C_SUB_SENTENCE,sChar)||strstr(SEPERATOR_E_SENTENCE,sChar)||strstr(SEPERATOR_E_SUB_SENTENCE,sChar)||strstr(SEPERATOR_LINK,sChar))
{//Reach end of a sentence.Get a whole sentence
if(!strstr(SEPERATOR_LINK,sChar))//Not link seperator
{
strcat(sSentence,sChar);
}
if(sSentence[0]!=0&&strcmp(sSentence,SENTENCE_BEGIN)!=0)
{
if(!strstr(SEPERATOR_C_SUB_SENTENCE,sChar)&&!strstr(SEPERATOR_E_SUB_SENTENCE,sChar))
strcat(sSentence,SENTENCE_END);//Add sentence ending flag
Processing(sSentence,1);//Processing and output the result of current sentence.
Output(m_pResult[0],sSentenceResult,bFirstIgnore);//Output to the imediate result
//bFirstIgnore=true;
strcat(sResult,sSentenceResult);//Store in the result buffer
}
if(strstr(SEPERATOR_LINK,sChar))//Link the result with the SEPERATOR_LINK
{
strcat(sResult,sChar);
strcpy(sSentence,SENTENCE_BEGIN);//Add a sentence begin flag
//sSentence[0]=0;//New sentence, and begin new segmentation
//bFirstIgnore=false;
}
else if(strstr(SEPERATOR_C_SENTENCE,sChar)||strstr(SEPERATOR_E_SENTENCE,sChar))
{
strcpy(sSentence,SENTENCE_BEGIN);//Add a sentence begin flag
//sSentence[0]=0;//New sentence, and begin new segmentation
//bFirstIgnore=false;
}
else
{
strcpy(sSentence,sChar);//reset current sentence, and add the previous end at begin position
}
}
else //Other chars and store in the sentence buffer
strcat(sSentence,sChar);
}
if(sSentence[0]!=0&&strcmp(sSentence,SENTENCE_BEGIN)!=0)
{
strcat(sSentence,SENTENCE_END);//Add sentence ending flag
Processing(sSentence,1);//Processing and output the result of current sentence.
Output(m_pResult[0],sSentenceResult,bFirstIgnore);//Output to the imediate result
strcat(sResult,sSentenceResult);//Store in the result buffer
}
delete [] sSentence;//FREE sentence buffer
delete [] sSentenceResult;//free buffer
return true;
}
//Adjust the result with some rules
bool CResult::Adjust(PWORD_RESULT pItem,PWORD_RESULT pItemRet)
{
int i=0,j=0;
unsigned int nLen;
char sSurName[10],sSurName2[10],sGivenName[10];
bool bProcessed=false;//Have been processed
while(pItem[i].sWord[0]!=0)
{
nLen=strlen(pItem[i].sWord);
bProcessed=false;
//Rule1: adjust person name
if(pItem[i].nHandle==28274&&ChineseNameSplit(pItem[i].sWord,sSurName,sSurName2,sGivenName,m_uPerson.m_dict)&&strcmp(pItem[i].sWord,"叶利钦")!=0)//'nr'
{//Divide name into surname and given name
if(sSurName[0])
{
strcpy(pItemRet[j].sWord,sSurName);
pItemRet[j++].nHandle=28274;
}
if(sSurName2[0])
{
strcpy(pItemRet[j].sWord,sSurName2);
pItemRet[j++].nHandle=28274;
}
if(sGivenName[0])
{
strcpy(pItemRet[j].sWord,sGivenName);
pItemRet[j++].nHandle=28274;
}
bProcessed=true;
}
//Rule2 for overlap words ABB 一段段、一片片
else if(pItem[i].nHandle==27904&&strlen(pItem[i+1].sWord)==2&&strcmp(pItem[i+1].sWord,pItem[i+2].sWord)==0)
{//(pItem[i+1].nHandle/256=='q'||pItem[i+1].nHandle/256=='a')&&
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
strcat(pItemRet[j].sWord,pItem[i+2].sWord);
pItemRet[j].nHandle=27904;
j+=1;
i+=2;
bProcessed=true;
}
//Rule3 for overlap words AA
else if(nLen==2&&strcmp(pItem[i].sWord,pItem[i+1].sWord)==0)
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
//24832=='a'*256
pItemRet[j].nHandle=24832;//a
if(pItem[i].nHandle/256=='v'||pItem[i+1].nHandle/256=='v')//30208='v'8256
{
pItemRet[j].nHandle=30208;
}
if(pItem[i].nHandle/256=='n'||pItem[i+1].nHandle/256=='n')//30208='v'8256
{
pItemRet[j].nHandle='n'*256;
}
i+=1;
if(strlen(pItem[i+1].sWord)==2)
{//AAB:洗/洗/脸、蒙蒙亮
if((pItemRet[j].nHandle==30208&&pItem[i+1].nHandle/256=='n')||
(pItemRet[j].nHandle==24832&&pItem[i+1].nHandle/256=='a')
)
{
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
i+=1;
}
}
j+=1;
bProcessed=true;
}
//Rule 4: AAB 洗/洗澡
else if(nLen==2&&strncmp(pItem[i].sWord,pItem[i+1].sWord,2)==0&&strlen(pItem[i+1].sWord)==4&&(pItem[i].nHandle/256=='v'||pItem[i].nHandle==24832))//v,a
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
//24832=='a'*256
pItemRet[j].nHandle=24832;//'a'
if(pItem[i].nHandle/256=='v'||pItem[i+1].nHandle/256=='v')//30208='v'8256
{
pItemRet[j].nHandle=30208;
}
i+=1;
j+=1;
bProcessed=true;
}
else if(pItem[i].nHandle/256=='u'&&pItem[i].nHandle%256)//uj,ud,uv,uz,ul,ug->u
pItem[i].nHandle='u'*256;
else if(nLen==2&&strncmp(pItem[i].sWord,pItem[i+1].sWord,2)==0&&strlen(pItem[i+1].sWord)==4&&strncmp(pItem[i+1].sWord+2,pItem[i+2].sWord,2)==0)
{//AABB 朴朴素素 枝枝叶叶
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
strcat(pItemRet[j].sWord,pItem[i+2].sWord);
pItemRet[j].nHandle=pItem[i+1].nHandle;
i+=2;
j+=1;
bProcessed=true;
}
else if(pItem[i].nHandle==28275)//PostFix
{
if(m_uPlace.m_dict.IsExist(pItem[i+1].sWord,4))
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
pItemRet[j].nHandle=28275;
i+=1;
j+=1;
bProcessed=true;
}
else if(strlen(pItem[i+1].sWord)==2&&CC_Find("队",pItem[i+1].sWord))
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
pItemRet[j].nHandle=28276;
i+=1;
j+=1;
bProcessed=true;
}
else if(strlen(pItem[i+1].sWord)==2&&CC_Find("语文字杯",pItem[i+1].sWord))
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
pItemRet[j].nHandle=28282;
i+=1;
j+=1;
bProcessed=true;
}
else if(strlen(pItem[i+1].sWord)==2&&CC_Find("裔",pItem[i+1].sWord))
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
pItemRet[j].nHandle=28160;
i+=1;
j+=1;
bProcessed=true;
}
}
else if(pItem[i].nHandle==30208||pItem[i].nHandle==28160)//v
{
if(strlen(pItem[i+1].sWord)==2&&CC_Find("员",pItem[i+1].sWord))
{
strcpy(pItemRet[j].sWord,pItem[i].sWord);
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
pItemRet[j].nHandle=28160;
i+=1;
j+=1;
bProcessed=true;
}
}
else if(pItem[i].nHandle==28280)
{//www/nx ./w sina/nx; ＥＩＭ/nx -６０１/m
strcpy(pItemRet[j].sWord,pItem[i].sWord);
pItemRet[j].nHandle=28280;
while(pItem[i+1].nHandle==28280||strstr(".．",pItem[i+1].sWord)||(pItem[i+1].nHandle==27904&&IsAllNum((unsigned char *)pItem[i+1].sWord)))
{
strcat(pItemRet[j].sWord,pItem[i+1].sWord);
i+=1;
}
j+=1;
bProcessed=true;
}
if(!bProcessed)
{//If not processed,that's mean: not need to adjust;
//just copy to the final result
strcpy(pItemRet[j].sWord,pItem[i].sWord);
pItemRet[j++].nHandle=pItem[i].nHandle;
}
i++;
}
pItemRet[j].sWord[0]=0;//Set ending
return true;
}
//Generate Word according the segmentation route
bool CSegment::GenerateWord(int **nSegRoute, int nIndex)
{
unsigned int i=0,k=0;
int j,nStartVertex,nEndVertex,nPOS;
char sAtom[WORD_MAXLENGTH],sNumCandidate[100];
ELEMENT_TYPE fValue;
while(nSegRoute[nIndex][i]!=-1&&nSegRoute[nIndex][i+1]!=-1&&nSegRoute[nIndex][i]<nSegRoute[nIndex][i+1])
{
nStartVertex=nSegRoute[nIndex][i];
j=nStartVertex;//Set the start vertex
nEndVertex=nSegRoute[nIndex][i+1];//Set the end vertex
nPOS=0;
m_graphSeg.m_segGraph.GetElementInfo(nStartVertex,nEndVertex,fValue,nPOS);
sAtom[0]=0;
while(j < nEndVertex)
{
//Generate the word according the segmentation route
strcat(sAtom,m_graphSeg.m_sAtom[j]);
j++;
}
m_pWordSeg[nIndex][k].sWord[0]=0;//Init the result ending
strcpy(sNumCandidate,sAtom);
while(sAtom[0]!=0 && (IsAllNum((unsigned char *)sNumCandidate)||IsAllChineseNum(sNumCandidate)))
{
//Merge all seperate continue num into one number
//sAtom[0]!=0: add in 2002-5-9
strcpy(m_pWordSeg[nIndex][k].sWord,sNumCandidate);
//Save them in the result segmentation
i++;//Skip to next atom now
sAtom[0]=0;
while(j<nSegRoute[nIndex][i+1])
{//Generate the word according the segmentation route
strcat(sAtom,m_graphSeg.m_sAtom[j]);
j++;
}
strcat(sNumCandidate,sAtom);
}
if(m_pWordSeg[nIndex][k].sWord[0]==0)//Have never entering the while loop
{
strcpy(m_pWordSeg[nIndex][k].sWord,sAtom);
//Save them in the result segmentation
}
else
{//It is a num
if(strcmp("－－",m_pWordSeg[nIndex][k].sWord)==0||strcmp("—",m_pWordSeg[nIndex][k].sWord)==0||m_pWordSeg[nIndex][k].sWord[0]=='-')//The delimiter "－－"
{
nPOS=30464;//'w'*256;Set the POS with 'w'
i--;//Not num, back to previous word
}
else
{//Adding time suffix
char sInitChar[3];
unsigned int nCharIndex=0;//Get first char
sInitChar[nCharIndex]=m_pWordSeg[nIndex][k].sWord[nCharIndex];
if(sInitChar[nCharIndex]<0)
{
nCharIndex+=1;
sInitChar[nCharIndex]=m_pWordSeg[nIndex][k].sWord[nCharIndex];
}
nCharIndex+=1;
sInitChar[nCharIndex]='\0';
if(k>0&&m_pWordSeg[nIndex][k-1].nHandle==27904&&(strcmp(sInitChar,"—")==0||sInitChar[0]=='-')&&(strlen(m_pWordSeg[nIndex][k].sWord)>nCharIndex))
{//3-4月 //27904='m'*256
//Split the sInitChar from the original word
strcpy(m_pWordSeg[nIndex][k+1].sWord,m_pWordSeg[nIndex][k].sWord+nCharIndex);
m_pWordSeg[nIndex][k+1].dValue=m_pWordSeg[nIndex][k].dValue;
m_pWordSeg[nIndex][k+1].nHandle=27904;
m_pWordSeg[nIndex][k].sWord[nCharIndex]=0;
m_pWordSeg[nIndex][k].dValue=0;
m_pWordSeg[nIndex][k].nHandle=30464;//'w'*256;
m_graphOptimum.SetElement(nStartVertex,nStartVertex+1,m_pWordSeg[nIndex][k].dValue,m_pWordSeg[nIndex][k].nHandle);
nStartVertex+=1;
k+=1;
}
unsigned int nLen=strlen(m_pWordSeg[nIndex][k].sWord);
if((strlen(sAtom)==2&&CC_Find("月日时分秒",sAtom))||strcmp(sAtom,"月份")==0)
{//2001年
strcat(m_pWordSeg[nIndex][k].sWord,sAtom);
nPOS=29696;//'t'*256;//Set the POS with 'm'
}
else if(strcmp(sAtom,"年")==0)
{
if(IsYearTime(m_pWordSeg[nIndex][k].sWord))//strncmp(sAtom,"年",2)==0&&
{//1998年，
strcat(m_pWordSeg[nIndex][k++].sWord,sAtom);
nPOS='t'*256;//Set the POS with 'm'
}
else
i--;//Can not be a time word
}
else
{
//早晨/t 五点/t
if(strcmp(m_pWordSeg[nIndex][k].sWord+strlen(m_pWordSeg[nIndex][k].sWord)-2,"点")==0)
{
nPOS='t'*256;//Set the POS with 'm'
}
else
{
if(m_pWordSeg[nIndex][k].sWord[0]!='.')
nPOS='m'*256;//Set the POS with 'm'
if(nLen>1&&m_pWordSeg[nIndex][k].sWord[nLen-1]=='.')
{//Get rid of . example 1.
m_pWordSeg[nIndex][k].sWord[nLen-1]=0;
i--;
}
}
i--;//Not num, back to previous word
}
}
fValue=0;
nEndVertex=nSegRoute[nIndex][i+1];//Ending POS changed to latter
}
m_pWordSeg[nIndex][k].nHandle=nPOS;//Get the POS of current word
m_pWordSeg[nIndex][k].dValue=fValue;//(int)(MAX_FREQUENCE*exp(-fValue));//Return the frequency of current word
m_graphOptimum.SetElement(nStartVertex,nEndVertex,fValue,nPOS);
//Generate optimum segmentation graph according the segmentation result
i++;//Skip to next atom
k++;//Accept next word
}
m_pWordSeg[nIndex][k].sWord[0]=0;
m_pWordSeg[nIndex][k].nHandle=-1;//Set ending
return true;
}